项目名称: 条件模型的计量经济学方法探讨及应用
项目编号: No.71301135
项目类型: 青年科学基金项目
立项/批准年度: 2014
项目学科: 管理科学
项目作者: 任宇
作者单位: 厦门大学
项目金额: 19万元
中文摘要: 在当前的经济学和金融学研究中,很多线性模型的参数值会随经济状态变化而变化。为了描述这种变化,研究者通常会预先设定一些函数来刻画模型中参数与经济状态的关系。这样,原有的线性模型就扩展为了条件模型 (conditional model)。然而,描述这些参数的函数在设定时往往带有较强的主观性,这种主观性会导致不同的学者在研究相同的经济学问题时,即使运用同样的数据来进行计量分析,也会得到不同的结论,甚至某些结论是相互排斥的。因此,我们提出新的计量经济学方法来正确地估计这些条件模型。本项目将非参数函数系数估计法与惩罚函数结合在一起来估计条件模型。其中,非参数函数系数估计法能避免错误模型设定所带来的误差,而惩罚函数则可以选择出最优的状态变量。在实证中,我们将运用这个新的理论来讨论条件资本资产定价模型以及风险资产的最优配置问题。
中文关键词: 条件模型;非参数估计;惩罚函数;资本资产模型;证券组合
英文摘要: In the current literature of economics and finance, the parameters in the linear models are varying with economic conditions. In order to capture this variation, researchers usually set up some functions between the parameters and the economic conditions. Therefore, the linear model can be extended into the conditional model. But, it is quite arbitrary to choose the functions, which may lead to the situation that different researchers may get different results even though they are studying the same economic phenomena and using the same data. Sometimes, their results contradict with each other. So, we propose a new econometric method to estimate the conditional models. Particularly, we use the nonparametric estimation combined with penalty function to fulfill this job. Nonparametric estimation can avoid the error due to model misspecification, and penalty function can help us choose optimal state variables. Empirically, we use our new method to study the conditional capital asset pricing model and portfolio allocation.
英文关键词: Conditional model;Nonparametric estimation;Penalty function;CAPM;Portfolio Selection